Deep Research Workflow FAQs

Question 1

Is the research data stored locally?

Accepted Answer

Yes, all artifacts, including raw data, intermediate logs, and the final report, are stored in a semantic directory structure under .research/ for full auditability.

Question 2

Does this skill provide answers directly in the chat?

Accepted Answer

No, the skill is designed to prioritize 'File Delivery.' It generates a standalone, structured Markdown report in a dedicated research directory to maintain professional formatting and depth.

Question 3

How does the Deep Research skill handle web access?

Accepted Answer

It follows a strict priority: first using installed skills, then MCP tools like Firecrawl or Tavily, and finally falling back to shell-based networking only when explicitly authorized.

Question 4

Can I run multiple research tasks at the same time?

Accepted Answer

Yes, it generates a scheduling script that uses parallel sub-processes to handle multiple sub-goals simultaneously, significantly reducing the time required for broad research.

Question 5

What happens if a sub-task fails during research?

Accepted Answer

The skill captures exit codes and logs for every sub-process. It provides a detailed error report and allows for retries or manual intervention without restarting the entire workflow.

Deep Research Workflow

Key Features

Use Cases

Deep Research Workflow

Key Features

Use Cases